Habitat mapping for vulnerable
species is critical to assessing reserve design alternatives in terms of
conservation goals. Habitat
distributions for many of the vulnerable species are poorly known and published
accounts of known populations are few; therefore, habitat modeling based on
environmental characteristics was conducted in order to provide the most
complete, scientifically based depiction of species habitat. Recognizing the critical knowledge of
many Pima County biologists, these “expert reviewers” were asked to be part of
the modeling process.
Reviewers identified key environmental variables describing habitat and
helped GIS analysts score environmental characteristics for each species. Analysts then built GIS models based on
these environmental parameters resulting in maps of high, medium, and low
potential habitat. Biologists were
then asked to review habitat maps and revise model parameters if necessary. This iterative process of GIS analysis
and biological review resulted in refined models that more closely represented
vulnerable species habitat.
During this process, data gaps
were identified and filled with new and updated mapping for most environmental
variables. After model parameters
were finalized and the GIS approach refined, remaining errors in habitat
distribution were corrected manually by reviewers. For example, biogeographical ranges were delineated and
applied to final models to reduce the extent of predicted habitat to more
closely represent the true potential range of the species. This final step improved the accuracy
of habitat maps.
Although
mapping potential species habitat is critical to reserve design and assessment,
mapping locations of known populations and critical areas for conservation are
also important. Therefore, HDMS,
the most extensive published database of species known locations, was used in
mapping and conservation goals assessment. This database was continually augmented with locations
mapped by species experts throughout the map review process. Expert reviewers, coordinated by SRG,
were also asked to delineate and describe priority areas for conservation for
each vulnerable species. Priority
conservation area 1 designates areas that experts believe are critical to
include in the reserve system, priority conservation area 2 designates areas
that are of value to the reserve system, priority conservation area 3
designates areas important for habitat connectivity, and priority conservation
area 4 designates areas with potential for habitat restoration. A thorough discussion of this effort is
included in the forthcoming report, “Priority Conservation Areas”, June
2001. Priority conservation areas
and all known locations are shown together with modeled habitat in each species
account.
GIS was used to generate habitat models for species by summarizing key environmental characteristics that comprise species habitat. After identifying important environmental variables, each environmental characteristic was scored as potential habitat for each species. These species-habitat scores are compiled in a species-environment matrix where each characteristic (i.e., mixed grass-scrub) of a variable (i.e., vegetation) is valued from low to high (1-3) for each species. Environmental characteristics that have no value to species are scored 0, and characteristics which act as barriers to habitat are scored as “MASK”, which means the areas they cover are excluded from potential habitat in the model.
Variables used in the models are
vegetation/land cover, plus urban, meso-riparian and xero-riparian, perennial
and intermittent streams, shallow groundwater, springs, elevation, slope,
aspect, landform, cave/mine potential, geology, and soils. A total of 115 characteristics for
these 15 variables were scored as potential habitat for each species. Characteristics of these variables are
well understood for some species (such as a fish requiring a perennial stream),
but many are not. In some cases, a
“best guess” was recorded in the table cells of the species-environment
matrix. All scores have been
reviewed and revised by species experts.
The species-environment matrix
then served to define model parameters for GIS grid analysis. In the GIS, each variable exists as a
data layer comprised of regular 100 m grid cells covering the entire
county. New grids are created for
each species, with grid cell values storing habitat scores from the
species-environment matrix. In the
GIS modeling process, grids that comprise habitat for a species are “stacked”
on top of one another. Grid cell
stacks are then added to create a new grid that represents the sum of all
environmental characteristics for a species. This grid is the habitat model.
The quality of resulting GIS
models depends on both the accuracy and resolution of environmental variable
mapping and the appropriate scoring of environmental characteristics for each
species. The advantage of using
GIS is that it is an unbiased, systematic approach to synthesizing large
amounts of information.
Furthermore, all model parameters are explicit which facilitates review
and revision. Multiple updates to
GIS data layers for environmental variables and species scores in the
species-environment matrix have been made during the last 12 months and are
ongoing, so models will continue to improve.